Featured Post

This is the Kodak Moment for the Auto Industry

Plug-In Drivers Not Missin' the Piston Electric vehicles are here to stay. Their market acceptance and growth will continue. Why?...

Monday, May 25, 2020

What is Tesla's Project Dojo?

Tesla has made significant investments in artificial intelligence (AI). AI is the key to Tesla's full self-driving (FSD) future. Yet, Elon Musk has also called AI humanity's “biggest existential threat.” How do you reconcile this dichotomy? The answer is simple, Narrow AI vs General AI. A narrow AI is trained for a particular task such as playing a particular game or language processing. These narrow intelligences are not transferable. A narrow chess AI will not know anything about checkers despite the two games sharing a board. Whereas, a General AI (sometimes called Strong AI or Artificial General Intelligence(AGI)) is the hypothetical ability of a system to learn any intellectual task that a human could learn. Skills an AGI learned in one arena could be applied in new areas and an artificial superintelligence could quickly develop. An artificial superintelligence may find humans are irrelevant or worse, a threat. This is the “existential threat” that concerns Musk. 

So Tesla's FSD system will be a narrow AI, able to drive your car and you'll even be able to tell it where you'd like to go. You won't, however, be able to chat with the FSD AI about your day, but at least you'll know it won't decide that the best way to reduce traffic accidents is to kill all humans. 

Tesla's AI investments to date include creating an AI software development and validation team, creating a data labeling team, and creating an FSD hardware team to design their own custom neural network inference engine. Next on Tesla's AI investment list is "Project Dojo."

Project Dojo

We've been given a few hints about Dojo: Musk talked about it in the 2019 financial call and Tesla's Director of Artificial Intelligence and Autopilot, Andrej Karpathy, has talked about it at multiple AI conferences. We'll discuss how neural nets work and then move into some wild speculation; but first, we have to acknowledge the Dad Joke that is the name Project Dojo. We know that Project Dojo is intended to vastly improve the Autopilot Neural Network training. If you want to train, where do you go? A Dojo, of course. 

Before we get into Dojo we need to cover a few basics about neural networks. There are two fundamental phases to neural networks (NN): Training and Inference.


NNs have to be trained. Training is a massive undertaking. This is when the digital ocean of data that is the training dataset must be digested. It takes terabytes of data and exaflops of compute to train a complex NN. Through training the NN forms "weights" for nodes. When the training is complete, the resulting NN is tested. A test dataset that was not part of the training dataset, where the expected results are known, is thrown at the resulting network and if the NN is properly trained, it infers the correct answer for each test. Since Project Dojo is all about training, we'll dig more into this later. Depending on the use case, there may be several stages of simulation and testing before the NN is deployed. Deploying the NN leads us to our next phase, Inference.


When a neural network receives input, it infers things about the input based on its training; this is known as “inference.” These inferences may or may not be correct. Compared to training, the storage and compute power needed for inference is significantly lower. However, in real-time applications, the inference needs to happen within milliseconds; whereas training can take hours, days, or weeks.

Unlike training, inference doesn't modify the neural network based on the results. So when the NN makes a mistake, it is important that these are captured and fed back to the training phase. This brings us to a third (optional) phase, Feedback.


You may have heard the phrase "Data is the new Oil." Nowhere is this more applicable than AI training datasets. If you want an AI that performs well, you have to give it a training set that covers many examples of all of the types of situations that it may encounter. After you have deployed the AI, you have to collect the situations where it did the wrong thing, label it with the expected result, and add this (and perhaps hundreds or thousands of examples like it) to the training dataset. This allows the AI to iteratively improve. However, it means that your training dataset grows with each iteration and so does the amount of computing horsepower needed for training.

Tesla's Autopilot Flywheel 

Now that we've ever so briefly covered AI basics, let's look at how these apply to Tesla's FSD.

Let's start with Deploying the Neural Net. Every car that Tesla makes today is a connected car that receives over-the-air updates. This allows the cars to receive new software versions frequently. When a new version of Autopilot is deployed, Tesla collects data about its performance. The AI makes predictions such as the path of travel, where to stop, et cetera. If Autopilot is driving and you disengage it, this may be because it was doing something incorrectly. These disengagements are reported back to Tesla (assuming you have data sharing enabled). The report could be a small file that only has the data labels and a few details or it could be streams of sensor data and clips of video footage depending on the type of disengagement and the types of situations that Tesla is currently adding to their training set.

Even if Autopilot is not engaged, it is running in "shadow mode." In shadow mode, it is still making predictions and taking note when you, the human driver, don't follow those predictions. For example, if it predicts that the road bends to the left, but you go straight, this would be noted and potentially reported back to the mothership. If Autopilot infers that a traffic light is green but you stop, this data would again likely be noted and potentially reported back.

Tesla has about a million vehicles on the road today collectively driving about 15 billion miles each year. The bulk of these cars are from Tesla's Fremont factory. Tesla now has a second factory, Giga Shanghai, putting cars on the road. Soon Giga Berlin and Giga Austin (or will it be Tulsa?) will join them. All of this will result in a large amount of data for the training dataset.

The bigger the training set, the longer it takes to process. However, with a system like this, the best way to improve it is to quickly iterate (deploy it, collect errors, improve, repeat). If training takes months, this slows down the flywheel. How do you resolve this? With a supercomputer dedicated to AI training. This is Project Dojo: make a training system that can drink in the oceans of data and produce a trained NN in days instead of months.

A Cerebras Wafer Scale Engine


At the start, I promised some speculation. As promised, here it is.

The size of the chips used for AI training has been increasing every year. From 2013 to 2019, AI chips increased by about 50% in size. A startup called Cerebras saw this trend and extrapolated it to its natural conclusion of 1 chip per wafer. For comparison, the Cerebras chip is 56 times bigger than the largest GPU made in 2019, it has 3,000 times more on-chip memory, and it has more than 10,000 times the memory bandwidth.

This wafer-scale chip is an AI training accelerator and my conjecture is that a Cerebras chip will be at the heart of Project Dojo. This wafer-scale chip is the biggest (literally and figuratively) breakthrough in AI chip design in a long time.

There is one (albeit tenuous) thread that connects Tesla and Cerebras, both are part of ARK Invest's disruption portfolio. ARK has investments in both companies and meets with their management teams. When there are two companies that could mutually benefit working together and it would benefit their mutual investor, ARK, you can bet that introductions would be made.

Thursday, January 16, 2020

10 Years of Trading Tesla (TSLA)

Tesla's stock has been on a tear recently. I've been buying (and occasionally selling) the stock since its IPO in 2010. Below is a brief history of my trading activity.

Of course, I have no way of knowing what the stock will do tomorrow, so don't take this as stock advice.

I bought my first shares soon after the IPO. The stock opened at about $20 and had a dip over the next few weeks. In late June and early July of 2010, I bought at $18, $17.84, and (the best price I picked some up was) $16.01 per share.

I held these shares for nearly 6 years, until early 2016. Why did I sell them then? Two reasons. First, after a stock has had a good run (from $18 to $249 (or ~1400%) in this case), I like to take out my initial stake so that no matter what happens to the stock after that, I will always be net positive. The second reason I sold was that we were going to buy a new car in 2016. I didn't sell all of my shares.

My timing to sell was great. The stock dipped later in 2016 and I was able to buy the shares back at a lower price.

After taking delivery of the car in the fall of 2016, my view of the company changed. This was not my first EV (it was my 3rd actually). I knew that EVs were the future of personal transportation, but Tesla was lightyears ahead of everyone else. There was no other car that could compare. After owning a Tesla, all other cars (electric or not) seem like relics from a bygone era. They did unlock as you walked up to them, you had to push a button or turn a key to start it and stop it, they had tiny screens, they didn't have vast free Supercharging networks, they didn't have 200+ miles of range, they didn't receive firmware updates over-the-air...

Based on this two-pronged belief (1: EVs are the future. 2: Only Tesla has cracked the code), throughout 2017 and 2018, I was buying TSLA whenever the price dropped below $300. At the end of 2018, I sold a portion of my shares at $375. The reason we sold this time was once again, to buy a Tesla.

Again my sell timing was lucky. We sold near a local maximum. Soon after we sold, the SEC became concerned with Musk's infamous 420 tweet. This, and other concerns, drove the stock price down in the first half of 2019. This allowed me to buy shares back in the $200s, I even picked up some in May of 2019 for $185 per share. I had just sold for $375 and now I was able to buy it at half that price. How great is that? I understand that an investor would not be happy if they had bought at $375 and saw their investment halved. I, on the other hand, was convinced that this slump in the stock price was temporary. Issues like this get resolved and Tesla still made the best vehicles in a fast-growing category.

Now, it's early 2020 and the stock is over $500 per share. Again, I am taking some profits for the same 2 reasons I did initially. One, to remove my seed funds. Doing this allows me to sleep soundly at night. TSLA is a volatile stock. If it goes up, I still own shares and I'll share in the rewards. But if it goes down, I'm not concerned. By removing the money I initially put into it (plus a little), I am guaranteed, that (even if the stock goes to zero) I've made money on my Tesla trades. And the second reason is to again buy a Tesla product. This time we are getting Powerwalls installed on our home. More on that in later posts.

It only seems right that after making money on their stock that I should share the profits with them by buying their products. I've certainly done the same with Amazon, Netflix, and Google.

I'm still holding TSLA, I'm long the stock.


Sunday, September 15, 2019

3 Years of Tesla Model X Ownership

In September of 2016, I bought a Tesla Model X 90D. This has been my daily driver ever since and we've taken it on multiple road trips. It has performed flawlessly. Below, we'll look at road trips, fuel costs, upgrades, and battery degradation during these years of ownership.

3-year-old (2016) Tesla Model X fresh and clean after a wash

Mileage and Road Trips

From our homebase in Portland, OR we've driven to Grants Pass; eastern Oregon; Bend, OR; San Diego, CA; Great Wolf Lodge; the dunes of Florence; Thor's Well; Crater Lake; Oregon Wildlife Safari; Cove Palisades; The Oregon Caves, and many other destinations.

On this 3rd anniversary of ownership, the X has 32,669 miles on the o-meter.

The Model X is a great vehicle for road trips. Around here, the Tesla charging network makes it easy to recharge and the stop time is just right to stretch your legs and grab a snack. Plenty of hotels have chargers, so you can start out each day with a full charge.

Fuel Cost

Can you call electricity a "fuel"? Either way, here's how much it cost us to drive these 32k miles.

About 8,000 of these miles were with free Supercharging. The bulk of the remaining miles were charged up at home in our garage. We have the time-of-use plan with our local utility and we are only charged 4.209¢ per kWh during overnight off-peak hours.

Doing a little math, we've paid $634 for 24,668 miles of travel, or 32,668 miles if you include those fueled by Superchargers. $634 for 3 years of driving is pretty good, but how does that compare to the cost we'd have paid to fuel a gas vehicle?

For the comparison, we'll look at two other Luxury Midsize SUVs from the same year: a 2016 BMW X5 M AWD 4DR and a 2016 Porsche Cayenne Turbo S. These get 14 city /19 hwy and 14/21 MPG respectively. Generously assuming the gas in the tank from the dealership covered the 668 miles, that leaves 32,000 miles worth of gas to buy.

The BWM X5 would burn about 1,940 gallons at a cost of $4,975 to travel 32k miles. The Porsche Cayenne Turbo S is only slightly better, burning 1,830 gallons of gas at a cost of $4,690. 

The Model X cost only 13.5% of the cost of the Porsche Cayenne Turbo S to fuel. That's as if we were paying 35¢ per gallon. When's the last time gasoline was 35¢ a gallon? It was around the time that Neil Armstrong walked on the Moon and that price wasn't going to last long because the OPEC oil embargo would soon follow.

As I type this, the big news story is "Two Major Saudi Oil Installations Hit by Drone Strike." Gasoline has been a problem for my entire lifetime, I think it's time to move on from this dysfunctional relationship. Electricity prices are far less volatile. No country has ever had their wind turbines as the target of a drone strike. 


One of the best features of any Tesla vehicle is the fact that it receives periodic software updates over-the-air. These updates add functionality and fun to the car. Here are a few of the things they've added during these 3 years:
  • Chill Mode
  • Easy Entry
  • Dog Mode 
  • Faster Supercharging
  • Battery Preconditioning 
  • New Application Launcher
  • Atari Games
  • Adaptive Suspension Damping Improvements
  • Driver Profile Key Linking
  • Heated Steering Wheel Improvements  
  • Sketchpad Improvements
  • Owners Manual Improvements
  • UI Improvements
  • Map Updates
Owners of newer Teslas might have noticed that I didn't list Navigate On Autopilot or some of the other AP related updates. That's because this was one of the last AP1 cars that Tesla made. I have no sour grapes over missing out on AP2+, I expect continuous innovation from Tesla.

Battery Degradation 

Long-time readers of this blog will know that I had a Nissan Leaf from 2011 until 2018. During those 7 years of ownership, I was greatly disappointed with the battery-life longevity. I wanted to keep the car for 10 years, but, for our needs, the battery range had degraded too much. So, battery degradation was one of my major concerns when I was EV shopping and this was one of the main reasons that I bought a Tesla. Taxi companies like Tesloop and others had bought Teslas and they were putting hundreds of thousands of miles on them each year. From their published data (see the chart below), the vehicles suffered about 10% of range loss over the first 100k miles and then the degradation flatted out and became negligible. So as long as you bought a car with ~10% more range than you needed, you should be fine.
Tesloop Battery Degradation Over 300k Miles

When new, our X had a 257-mile range. How has it held up? Today, it has a range of ~241 miles. That's a 6.3% range loss. In the first year, it lost 2.1% of range; in the second year, it lost an additional 2.2%; and in this third year, it lost 2.0%.

6.3% of Range Loss over 3 years
About the graph above, note that the left axis starts at 200 miles. This zoom-in allows you to see the degradation, but it also makes it look worse than it is. On most days, I'm only charging up to 160 miles, so the maximum range is not a significant factor.

Next year, I'm hoping to see less than 2% and for the degradation to flatten out at 230 miles. We shall see. If you plan on buying an EV and keeping it for a long time, make sure to account for some degradation as it ages.

Wrap Up

3 years of Model X ownership and I have no regrets. It was (is) the most I'd ever spent on a car. In fact, other than a house, it was the most expensive this I've ever purchased; and I'd do it again. Tesla's range, charging network, and fast charging time makes it so this can be your only vehicle. The battery management system smartly keeps the batteries from premature aging, so it should make it to the 10-year mark. I wonder what technomagical features the 2026 Tesla Model X will have.

Older Model X Reviews

You can see my 1-year review here and my 2-year review here.

Sunday, May 26, 2019

My First Tesla Grin

Nine years ago today, my friends, Jesse and Jim, and I went to a Tesla Roadster Ride & Drive event. There was no Tesla store in the Portland area at that time. The Seattle store was hosting a traveling roadshow and luckily we had appointments to the event.

I had no plans to buy a one hundred thousand dollar plus sports car, but I sure wanted to drive one. The car did not disappoint.

Getting in such a low-seated vehicle was an unusual feeling. At that time, my daily driver was an electric pickup truck. As I maneuvered the car to exit the parking lot, the smooth quiet motion of electric driving was familiar. Entering the road, I mashed the accelerator. Here's where the sportscar diverged from my electric truck. I was shoved back in my seat and a grin, assisted by G-forces, spread across my face.

I'd experienced the "EV grin" from the smooth quiet electric acceleration that feels like magical propulsion. The Tesla grin was the next level of this experience. As we merged onto the freeway, I was again able to mash the accelerator. This car was fun! There was no delay waiting for an engine to rev up before zipping ahead. As changed lanes (to the fast lane of course), I was again surprised by the car's response time. I was apparently used to sloppy comfort steering. This car had tight sports steering. It seemed to start the maneuver as soon as I thought about it.

I left that day excited for the future of EVs and knowing that I'd keep an eye on Tesla.

Six years later, at the same location, I attended the "Meet Model X" touring event with my friend Gary. Below is a photo of Gary being interviewed by the local media.

The media came along on Gary's Model X drive. The Falcon Wing door blew me away. This was far more practical than the Roadster and it even had towing capability. Soon after this event, I ordered a Model X.

Disclosure: I am long Tesla

Sunday, January 27, 2019

Alt Fuels Maps Compared

Energy.gov has a site that shows Alternative fueling stations. If you're interested in alt fuels, it is interesting to browse. For example, let's look at Propane, Hydrogen, & Electric:

Propane (LPG)
DC Fast Charging (CHAdeMO, Tesla, & CCS)
DC fast charging clearly has the most robust network among these three. Hydrogen only seems to be viable if you are in southern California or Silicon Valley and don't plan to take road trips.

The DC Fact charge map is perhaps not a fair map since there is no car out there today that could charge at all three types, so let's split these each out into their own maps.



Tesla Superchargers

The CCS coverage is primarily on the coasts and growing quickly. The CHAdeMO network is similar. It may be similar because many of the charging installations have both CCS and CHAdeMO stations.

The Tesla map has clusters in California and the northeast US, but it also has a spread of periodic stations along the major cross-country corridors. This is what's necessary in order to make long-distance driving liable for all but the most fervent.

Saturday, January 19, 2019

The Road To 2 Million EVs on US Roads

The modern EV era started with the introduction of the Chevy Volt and the Nissan Leaf in December of 2010. Now ~8 years later, we've had many notable milestones:
  • September 2015: Global plug-in sales passed 1 million units
  • December 2016: Global plug-in sales passed 2 million units
  • November 2017: Global plug-in sales passed 3 million units
  • December 2017: Annual global EV sales passed 1 million units
  • September 2018: Global plug-in sales passed 4 million units
  • September 2018: U.S. plug-in sales passed 1 million units
Looking at the final milestone on this list (1 million EVs in the US) has led to the question of when will the 2nd million EVs be on US roads. After reading a thread of bickering EV fans, I decided to make some forecast models of my own. I've made two models, one optimistic and the other less optimistic but still with strong growth.

The first million US EVs took 7 years and 10 months to sell. For the 2nd million, the pessimistic model forecasts ~3 years, whereas the optimistic model forecasts less than 2 years (~21 months). 

Each of these forecasts is shown in the charts below. Which one do you think is more accurate? Or do you have a different forecast? 

The next couple of years will be interesting for US EV sales. There are several headwinds and tailwinds that will impact EV sales.

Let's look at the headwinds first. The EV tax credit will phase out for Tesla, GM, and Nissan. GM has recently stopped making the Chevy Volt. As a vanguard for this new era, it is sad to see the Volt retired. The US economy is the next potential headwind. Janet Yellen, Alan Greenspan, & Morgan Stanley have all stated concerns about the economy slowing in 2019 and potentially entering a recession in 2020. These are key years in our path to 2 million EVs. An econmonic slowdown would certainly impact auto sales, EVs included.

As for tailwinds, many more EVs will be coming to market over the next two years including the $35k Tesla Model 3 and offerings from Audi, Kia, Mini, Porsche, Volvo, and others. New vehicles brings more options and more buyers. We can also expect to see redesigns of the Nissan Leaf, BMW i3, and the Tesla Model S & X. These refreshes should bring more range, faster charging, and/or other new features. These new and refreshed models will spur interest from a larger portion of the car buying market.

How these headwinds and tailwinds play out, will determine if the trend follows closer to the optimistic or the pessimistic forecast. Either way, it will be an exciting ride as personal transportation goes through its biggest transition in 100 years.